# ICT-FinD-Lab/alphagen

Generating sets of formulaic alpha (predictive) stock factors via reinforcement learning.

Repository: https://github.com/ICT-FinD-Lab/alphagen
Canonical: https://ross.abutalabs.com/products/alphagen
Language: Python
License Family: other
Topics: quantitative-trading, reinforcement-learning, symbolic-regression
Last push: 2026-06-04T05:54:38+00:00

## Health v2 (maintenance only)
Score: 70/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 85, release rhythm 35, longevity 100
- inputs: {"age_days": 1520, "days_push": 90, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases, no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1204, forks 322 (observed 2026-08-28T04:03:58.793307+00:00)

## What it is
AlphaGen is a Python research library that automatically generates formulaic alpha (predictive) stock factors using reinforcement learning, with optional LLM-based generation routines. It integrates with Microsoft Qlib for stock data preparation and supports adapters for custom alpha calculation pipelines.

## Use cases
- generate predictive stock factors with reinforcement learning
- mine formulaic alphas for quantitative trading strategies
- use LLMs to iteratively generate alpha expressions
- reproduce KDD 2023 alpha generation paper experiments
- integrate custom alpha calculation pipelines via adapters
- run symbolic regression on stock market data

## When to choose
- you need automated formulaic alpha mining for quant research
- you already use Qlib and want RL-based factor generation
- you want to experiment with LLM-driven alpha generation

## When to avoid
- you need production-ready trading infrastructure with execution and risk management
- you require a maintained, licensed dependency for commercial use (no license is specified)
- you need US/global market data out of the box (built-in pipeline targets Chinese A-shares via baostock)

## Facets
- artifact type: library
- maturity: active
- function: reinforcement-learning, machine-learning, llm-inference, data-science, trading
- domain: machine-learning, reinforcement-learning, fintech, data-science, large-language-models
- platform: python
- tags: quantitative-trading, symbolic-regression, alpha-mining, qlib, formulaic-alpha, research-code, linux, macos

## Member repositories
- ICT-FinD-Lab/alphagen (main) score 70

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:03:58.793307+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T06:19:37.811783+00:00, confidence not recorded.
  - readme: https://github.com/ICT-FinD-Lab/alphagen (fetched 2026-08-28T04:03:58.793307+00:00, sha a588c096ff63)
- Data as of 2026-08-30T08:39:29.467469+00:00.